MediBeat - HeartBeat Observation Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for MediBeat - HeartBeat Observation Trial (NCT04082819), a non-randomized na study. The primary outcome is Difference Between Systolic Manual Sphygmomanometer and Finger Pulse Oximeter Readings.

2. Clinical Question

3. Trial Design

The registry identifies this as a completed, non-randomized, double-masked, parallel-model study. Enrollment is listed as 94. The study began on 2019-10-04, and primary completion is listed as 2019-11-22. The record gives no phase designation beyond NA. SE Health is named as sponsor.

The study summary says that participants were recruited from Markham, Ontario and underwent alternating manual and device blood pressure measurements over a period of 45 minutes. Measurements were taken by trained nurses, and staff were consistent across the study duration. The primary outcome timeframe is recorded as 45 - 60 minutes. These are the two time descriptions in the record; it does not explain their relationship or identify a more specific schedule for individual readings.

Although the model is listed as parallel, the registry also describes blood pressure categories as low, medium, and high. The listed definitions are low systolic pressure from 0-129 and diastolic pressure from 0-79; medium systolic pressure from 130-160 and diastolic pressure from 80-100; and high systolic pressure from 161 or higher and diastolic pressure from 101 or higher. These are category descriptions, not evidence of randomized treatment arms. The record expressly indicates that the allocation was non-randomized and that the arms are not a like-for-like control comparison. The category-specific adverse-event listings also do not establish that participants were assigned to distinct intervention groups.

The record does not report the recruitment procedure in detail, criteria for entering a blood pressure category, allocation procedures, masking procedures, or how the device was configured. It also does not describe a randomization process. The design should therefore be read as a measurement-comparison study described by the registry, not as a randomized trial of competing treatments.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Analysis population for the primary outcome, as defined in the registry: The overall participants analyzed indicate the number of individuals who matched the ranges listed in the description for low, medium or high blood pressure categories, where they had blood pressure ranges falling between the respective values indicated for the systolic pressure measured.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Difference Between Systolic Manual Sphygmomanometer and Finger Pulse Oximeter Readings, measured over 45 - 60 minutes.

The primary endpoint is titled as the difference between systolic manual sphygmomanometer and finger pulse oximeter readings. Its unit is mmHg, and the summary parameter is the mean, with standard deviation listed as the dispersion measure. In general, a difference endpoint expresses how far apart paired measurement methods are on the measurement scale. The mean difference can describe the average direction and size of that separation, while the standard deviation describes variation in individual differences around the mean.

For each device reading, the record says that comparisons were made with the manual reading before it, R1, and the manual reading after it, R2. It then states that the reading with the best alignment was used. This choice is part of how the endpoint is defined: it is not described as a simple average of every device reading against every manual reading. The registry does not state the exact rule used to select the best-aligned reading or how the pairwise comparisons were combined into a participant-level result.

Comparing readings is useful when the question is whether a device produces measurements close to a reference method. General measurement principles distinguish average agreement from individual-level agreement: a small mean difference alone would not show how widely individual readings vary. The standard deviation is relevant to that variability, but the registry does not report its observed value. Nor does it report a confidence interval, limits of agreement, or another measure of the precision and spread of the device-reference differences. Consequently, the endpoint’s stated purpose is clear, but its numerical performance is not available in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean difference describe?

For a continuous measurement such as systolic blood pressure, a difference is formed by subtracting one method’s reading from the other according to a defined direction. The endpoint title identifies manual sphygmomanometer and finger pulse oximeter readings, but the registry does not specify the subtraction order in enough detail to establish which sign would mean that device readings were higher. It also does not report a mean difference. In general, the mean difference summarizes the average signed separation across the observations included in an analysis. Positive and negative differences can offset one another, so a mean near zero does not necessarily imply that individual measurements are close.

Why is dispersion relevant?

The registry names standard deviation as the dispersion measure for the mean endpoint. In general, standard deviation describes how much observed differences vary around their mean. It is not the same as the standard error, which describes uncertainty in an estimated mean, and neither quantity alone describes all features of agreement. A modest average difference could accompany a wide spread of individual differences. That is why interpreting a measurement-comparison result generally requires both a measure of average bias and information about variation. No endpoint standard deviation is reported in the registry record.

What does a confidence interval add?

In general, a confidence interval expresses the uncertainty around an estimate using a method with specified assumptions. For a mean difference, it can indicate the range of values compatible with the observed data and analysis. Its width depends on factors such as variability, the amount of usable information, and the method used to account for the study design. The registry does not report a confidence interval for the primary endpoint, nor does it provide the analysis details needed to reconstruct one. Without an estimate and its uncertainty, the precision of the study’s measured alignment cannot be evaluated.

How does pairing affect analysis?

Manual and device readings taken close together may be analyzed as paired observations, because each device measurement is associated with manual measurements around it. The record identifies the manual measurements as R1 and R2, with a device reading between them, and says that the best-aligned reading was used. In general, a paired analysis focuses on within-pair differences and can account for the relationship between paired measurements. The registry does not say whether the analysis used paired differences, how R1 and R2 were considered together, or how readings were aggregated for a participant. It does not report a model for repeated readings or explain how dependence among observations from the same participant was handled.

Why does the selection rule matter?

The instruction to use the reading with the best alignment is a meaningful part of the endpoint definition. In general, selecting among candidate readings based on closeness to a reference can make the resulting differences unlike those from a fixed, prespecified pairing. The exact impact depends on the selection process and analysis. The registry does not define the metric for best alignment, say whether selection was blinded to the result, or describe how often each candidate reading was selected. Those details would be important for interpreting the estimate and reproducing the analysis.

What is the role of over-dispersion?

Over-dispersion is a general issue in analyses where observations vary more than an assumed statistical model allows. It can lead to uncertainty estimates that are too small if the analysis treats related or variable observations as more independent than they are. For continuous repeated measurements, relevant concerns include within-participant correlation and heterogeneity in measurement differences. The registry does not report a model, variance assumptions, or any assessment of over-dispersion. It is therefore not possible to say whether this issue arose or how it was handled in this study.

What do the design labels imply?

The registry labels allocation as non-randomized, masking as double, and the model as parallel. These labels describe the recorded design, but they do not establish details beyond what the registry states. In general, non-randomized allocation does not support the same causal interpretation as random assignment, and a parallel-model label does not turn blood pressure categories into randomized groups. The record’s own caution is that the arms are not a like-for-like control comparison. For the measurement question, the important feature is the described comparison of manual and device readings; the record does not report a statistical method for evaluating their agreement.

What can be learned from planning assumptions?

Sample-size calculations are planning tools, not observed evidence about a study’s results. The stated planning assumptions describe a two-sided significance level of 0.05, equal allocation between arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. The record says the illustration is computed from the observed effect rather than the trial’s own design; it also distinguishes participant counts per arm from total event counts for time-to-event calculations. These assumptions do not constitute the trial’s sample-size justification, and the registry does not report that justification. A planning illustration cannot establish observed power or replace an estimate with its uncertainty. No numeric planning outputs are reported in the registry record.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Low Blood Pressure", "Medium Blood Pressure", "High Blood Pressure")
n   <- c(37, 54, 3)
m   <- c(3, 3, 11)
sp  <- c(6, 6, 10)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Low Blood Pressure  37  3.000  6.000  1.000  5.000
CARM  Medium Blood Pressure  54  3.000  6.000  1.362  4.638
CARM  High Blood Pressure  3  11.000  10.000  -13.841  35.841

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 2 outcome measures and 0 analyses. It identifies a primary systolic blood pressure difference endpoint, but does not describe the other outcome measure or provide an analysis plan. The number of outcomes alone does not establish whether multiplicity adjustments were needed or applied. In general, when several hypotheses are tested, the chance of at least one chance finding can increase; methods for handling this depend on the questions and analysis plan. The registry record does not report whether any adjustment was planned or used, nor whether any interim analyses occurred.

Missing-data methods are not reported in the registry record. It does not state whether every enrolled participant had usable paired readings, how incomplete measurement sequences were treated, or whether any observations or participants were excluded from analysis. Enrollment of 94 does not establish the size of the primary endpoint analysis population. The record also does not describe imputation, sensitivity analyses, or assumptions about missingness.

The population description refers to participants whose systolic measurements fell within the low, medium, or high ranges. The record does not give the counts in those categories or state whether the primary endpoint was analyzed across all eligible readings, separately within categories, or by another approach. It does not report a participant flow or a clear analysis population definition. These omissions limit assessment of which observations contributed to any result and how representative the analyzed measurements may have been. Since no endpoint results are posted in the registry record, neither multiplicity handling nor the effect of missing data can be assessed from numerical findings.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry lists serious adverse events separately for the low, medium, and high blood pressure categories. For each category, the reported number at risk for serious events is 94, and the number affected is 0. Deaths are listed as 0 for each category. Other adverse events are also listed with 94 at risk and 0 affected in each category. No top serious-event terms are listed.

These entries should be read as category-specific registry records, not added together. The categories are not described as mutually exclusive participant groups, and the repeated at-risk counts do not establish separate groups or additional participants. The registry record does not explain the relationship between the category listings and the total enrollment, nor does it provide event definitions, collection procedures, or a detailed safety analysis.

In general, a zero count in a registry listing means that no events are recorded in that entry; it does not by itself show that an event is impossible or that the study could rule out uncommon harms. Interpretation depends on the study’s safety surveillance and reporting methods, which are not reported. These safety figures are descriptive and do not establish the overall safety of a device or measurement approach.

13. What the Result Does, and Does Not, Mean

The registry record describes a study intended to assess how systolic readings from a finger pulse oximeter compare with manual sphygmomanometer measurements. It provides details about the measurement context, the use of readings before and after a device reading, and the selection of the reading with the best alignment. These details clarify the question the study set out to address.

The record does not report the primary endpoint estimate or its variability. It therefore does not show the average difference between measurement methods, the spread of individual differences, or the precision of an estimate. Without those results, the degree of agreement cannot be determined from the posted information. Nor does the record provide an analysis that would establish whether readings are sufficiently close for a particular purpose.

In general, measurement agreement and association are different ideas. Two methods can move in a similar direction while differing systematically, and a small average difference can conceal wide variation across individual measurements. Assessing agreement typically requires examining the differences and their distribution, along with uncertainty. The registry lists a mean and standard deviation as the endpoint summary, but supplies no observed values and reports no other agreement statistics.

The record also does not establish that the comparison applies beyond the study conditions. Participant characteristics, category sizes, detailed eligibility criteria, and measurement procedures are not fully reported. The non-randomized design and explicit warning that the arms are not like-for-like comparisons preclude treating the blood pressure categories as randomized controls. The measurements were taken by trained nurses, according to the study summary, but the record does not describe how results might vary with other operators or settings.

Accordingly, the study’s stated purpose can be understood, but its numerical findings cannot be interpreted from the registry record. The design description alone is not evidence that either measurement method is more accurate, and the available information does not support a conclusion about clinical performance or use.

14. Why This Trial Matters Statistically

Measurement studies address a statistical problem that differs from comparing clinical outcomes: whether methods produce sufficiently similar readings for the intended measurement context. The study’s focus on differences between manual sphygmomanometer and finger pulse oximeter systolic measurements is therefore relevant to biostatistical questions about pairing, variability, and agreement.

The record highlights why endpoint construction matters. It describes comparisons with manual readings before and after each device reading, then says the best-aligned reading was used. This selection rule can shape the distribution being summarized. Understanding the rule and its analysis is important for interpreting a mean difference and for determining whether another study could reproduce the same measurement procedure.

The design labels also require care. Non-randomized allocation and blood pressure category descriptions do not make the categories comparable intervention groups. The registry itself cautions that the arms are not a like-for-like control comparison. Keeping design labels separate from the actual measurement question prevents misleading causal interpretations.

Finally, the record illustrates the difference between describing an endpoint and reporting evidence about it. It identifies the unit, summary measure, dispersion measure, and timeframe, but does not give the mean difference, standard deviation, confidence interval, or analysis details. Without estimates and uncertainty, readers cannot evaluate the observed alignment or its precision. The statistical importance of this trial lies in the measurement question and the need for transparent endpoint and analysis reporting, while the posted registry information does not allow the result itself to be judged.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record leaves several trial-specific details unclear. The study summary describes alternating measurements over 45 minutes, while the primary outcome timeframe is listed as 45 - 60 minutes. No explanation reconciles these descriptions. The record does not specify the exact timing or number of readings, how the order of measurements was determined, or whether the measurement conditions were held constant beyond the statement that staff were consistent.

The primary endpoint title refers to finger pulse oximeter readings, while the listed device interventions name Contec CMS50EW and sphygmomanometer. The record does not explain the relationship between the named device and the HeartBeat algorithm, or how the algorithm generated the endpoint reading. It does not report device calibration, operating details, or the exact method for selecting the best-aligned comparison with R1 and R2. These omissions make the measurement process difficult to reproduce from the registry entry alone.

The registry does not report demographic or clinical characteristics, category-specific participant counts, eligibility criteria, or the number of participants with usable paired measurements. It does not explain whether the low, medium, and high categories were mutually exclusive or how category membership was determined when systolic and diastolic values might not fall into the same listed range. The number analyzed for the primary endpoint is not reported.

There is also no reported statistical analysis plan, model for repeated readings, confidence interval, or numerical endpoint result. The record does not describe missing-data handling, exclusions, or sensitivity analyses. These limitations prevent assessment of the magnitude, precision, and robustness of the measurement comparison.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: MediBeat - HeartBeat Observation Trial — NCT04082819.
• Registry record last updated 2020-04-01; status last verified 2020-03.
• Results first posted 2020-04-01.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry describes the completed MediBeat - HeartBeat Observation Trial, sponsored by SE Health, as a non-randomized, double-masked study with a parallel model. Enrollment is listed as 94. The study summary describes alternating blood pressure measurements using an experimental HeartBeat algorithm and manual sphygmomanometer readings, taken by trained nurses in Markham, Ontario. The primary endpoint is a mean difference in systolic readings, with standard deviation as the dispersion measure and mmHg as the unit.

The outcome description says device readings were compared with manual readings before and after them, and the best-aligned reading was used. The registry lists blood pressure categories but cautions that the arms are not a like-for-like control comparison. It reports no primary endpoint estimate, standard deviation, confidence interval, or analysis details. The numerical degree of alignment and its precision are therefore not reported in the registry record.